cosmosdb-design-patterns

Azure Cosmos DB Design Patterns

Architecture and integration patterns for Azure Cosmos DB applications, including AI agent orchestration with LangGraph.

When to Apply

Reference these guidelines when:

• Implementing materialized views with change feed
• Building ranking or leaderboard features
• Hydrating document references across containers
• Building LangGraph multi-agent applications with Cosmos DB
• Implementing human-in-the-loop flows
• Managing chat history and agent routing

Rules

• pattern-change-feed-materialized-views - Use Change Feed for cross-partition query optimization
• pattern-efficient-ranking - Efficient ranking approaches
• pattern-service-layer-relationships - Service layer for relationship hydration
• pattern-langgraph-multi-agent - StateGraph with conditional edges for multi-agent routing
• pattern-langgraph-interrupt-human - LangGraph interrupt for human-in-the-loop
• pattern-langgraph-resume-checkpoint - Resume from checkpoint after interrupt
• pattern-langgraph-agent-routing-cosmosdb - Persist active agent in Cosmos DB
• pattern-langgraph-fastapi-startup - Initialize LangGraph agents in FastAPI
• pattern-langgraph-chat-history-separate - Store chat history in dedicated container
• pattern-background-task-writes - FastAPI background tasks for non-blocking writes
• pattern-langgraph-async-cosmos-routing - Async Cosmos DB calls in LangGraph routing
• pattern-langgraph-async-cosmos-writes - Async active agent writes
• pattern-langgraph-agent-name-attribution - Tag AI messages with agent name
• pattern-ai-grounding-access - AI grounding access patterns

Full Compiled Document

For all rules expanded: AGENTS.md